Recent meta-analyses indicate that atypical antipsychotic medications are associated with increased mortality among patients with dementia. Moreover, our group has recently demonstrated that both atypical and conventional antipsychotic medications are associated with a greater mortality risk for elderly outpatients with dementia than other psychotropic medications often used for behavioral symptoms. These increased risks may persist for as long as 12 months following new antipsychotic starts. It is not yet known whether the mortality risk associated with the atypicals is an overall class effect, or associated with particular agents. There is also currently no information on causes of death from patients who die while taking these agents;this information would help researchers and clinicians to understand better the potential link between antipsychotics and mortality. Finally, no studies have been able to examine the impact of factors such as degree of cognitive impairment, type of behavior disorder or dementia stage on this relationship. Understanding the relationship between antipsychotics and mortality in the elderly is of profound importance to older patients as neuropsychiatric symptoms of dementia and other late-life psychotic syndromes are extremely common and problematic in these patients. In this study, we will extend our prior work to examine further critical questions including: 1) comparing mortality risks for individual antipsychotic agents in older patients both with and without dementia diagnoses in the 12 months following a new antipsychotic prescription;2) examining cause of death in patients who die in the 12 months following a new prescription for antipsychotic medications;and 3) investigating the impact of cognitive impairment, dementia stage and behavioral disturbance on medication preference and mortality. This will be a retrospective, cohort study using national Veterans Affairs registry and National Death Index data from fiscal years 2001-2004. The sample will be comprised of outpatients >65 years with new antipsychotic medication starts in FY 02 and 03. In the primary analysis, we will compare 12-month mortality rates using multivariate models, propensity- scoring methods, and instrumental variable methods. Secondary analyses will examine the impact of degree of cognitive impairment, dementia stage and neuropsychiatric symptoms on medication choice and mortality by collecting detailed medical chart data from a subset of patients. This will be the largest study to examine the relationship between antipsychotic agents and mortality in the elderly, and the only study to: 1) be sufficiently powered to examine the association between specific antipsychotic agents and mortality;2) address causes of death;and 3) consider explicitly the implications of cognitive and behavioral severity on risk of mortality with antipsychotics. The results of this study will be of immediate relevance to clinicians and policymakers.